Allow user to suppress NA warning in count variables

[TimTaylor]

As part of improvements for #113 we now warn if a count column contains NA values (and nudge users to handle these before calling incidence():

incidence2/R/incidence.R

Lines 368 to 375 in 88c1fb9

	nas_present <- sapply(.subset(DT,counts), anyNA)
	if (any(nas_present)) {
	missing_names <- names(nas_present)[nas_present]
	warnf(
	"`%s` contains NA values. Consider imputing these and calling `incidence()` again.",
	missing_names[1L]
	)
	}

We should provide arguments and options to suppress this if desired.